Introduction

Depression is ranked as one of the most prevalent disorders worldwide (together with headache disorders and low back pain) [1], affecting approximately 4.4% of the world population. Prevalence rates are also expected to rise, with depression projected to be the most prevalent disease by 2030 – due to aging populations, increasing mental health stigma, and rising socioeconomic stressors – underscoring the urgent need for novel interventions [2]. Left untreated, depression can also lead to increased comorbidities and worsening of somatic illnesses [3], as well as excess costs to health care systems [4]. Availability of effective treatments for depressive disorders are therefore of great importance. In order to understand the rationale behind these treatments, an understanding of the pathophysiology of depression itself is first necessary.

Depression is generally considered a mental/mood disorder, causing persistent feelings of sadness and loss of interest (see Table 1 for subclassifications), but the pathophysiology is still not fully understood. A complex interaction between neurotransmitters and receptor regulation/sensitivity have been suggested, implicating the role of serotonin [5] and noradrenalin/norepinephrine [6]. Likewise, exposure to chronic inflammation is thought to drive changes to the central nervous system (CNS) neurotransmitters, causing or worsening depression [7, 8].

Table 1 Classification of depression
Full size table

Existing treatment options for depression are multipronged, but can be broadly categorised into three groups. (1) Pharmaceuticals (e.g. selective serotonin reuptake inhibitors, SSRIs, selective serotonin norepinephrine reuptake inhibitors, SNRIs, and tricyclic antidepressants TCAs), that focus on increasing the levels of the aforementioned neurotransmitters in the CNS, by blocking their reuptake in the synaptic clefts [9]. (2) Talk therapy (including CBT and psychotherapy), focusing on the thought processes around depression [10]. (3) Exercise therapy, with aerobic group exercise having the strongest evidence to date [11]. The theory behind the efficacy of exercise therapy rests on several observed phenomena, just like antidepressants, physical activity significantly increases levels of noradrenaline/norepinephrine, among other neurotransmitters and hormones [12]. It also improves sleep quality [13] and lowers stress and inflammation, i.e. other possible causes of depression [7]. While all three therapeutic approaches have merit, and likely can be used together for synergic effects, this review will focus on the latter, and more specifically a new form of emerging exercise therapy: rock climbing/bouldering.

Several studies investigating the effects of exercise on depression have been carried out over the years, and later summarized in systematic reviews. The results of these systematic reviews have been somewhat mixed, but the majority support the use of exercise therapy. One systematic review [14], including 35 trials with 2,498 participants, investigating the effects of exercise therapy on major depression, found no significant effects. While another systematic review [15], including 39 trials with 2,326 participants, found a small effect, although not superior to psychotherapy or antidepressants. However, at least three systematic reviews [16,17,18], including up to 41 trials with up to 2,264 participants, one published only last year, all found that exercise therapy is effective in treating depression. One [18] also highlights that the “data strongly support the claim that exercise is an evidence-based treatment for depression”, while another [16] goes on to state that exercise therapy “should be offered as an evidence-based treatment” for depression. The aforementioned review also stresses the NICE-guidelines ((National Institutes for Health and Care Excellence, UK) [11], highlighting aerobic group exercise of moderate intensity.

As a new form of exercise therapy, rock climbing, has shown promise in improving both somatic conditions, such as low back pain [19,20,21], and mental conditions, such as anxiety disorders and attention deficit hyperactivity disorders (ADHD) [22]. Besides the pure physical and motor functional aspects of rock climbing, it is often carried out at high intensity, and differs from most other forms of exercise and sports in that it involves both a problem-solving aspect as well as exposure to considerable height, albeit protected with a rope or crash pad (see Table 2 for details). This element of height, and that climbs are often scaled with a partner – responsible for the safety of the climb – have been shown to build trust and confidence in the climber [23], i.e. feelings with an inverse relationship to depression [24]. It may be then, that this “adventure aspect” of rock climbing, has therapeutic effects beyond mere physical exercise, and may offer a unique therapeutic mechanism.

Table 2 Rock climbing
Full size table

Rock climbing has also been shown to improve cognitive abilities such as concentration, focus, problem-solving abilities, attention, and impulse control [22], faculties often associated with mindfulness, and too with an inverse relationship to depression [25]. If rock climbing could induce the same mindfulness-state, it may, for these reasons, offer greater therapeutic potentials than other forms of physical exercise. A recent study [26] supports this notion, suggesting that rock climbing could provide unique benefits by integrating physical, cognitive, and mindfulness elements, thereby potentially producing a synergistic effect on depressive symptoms.

Another obvious advantage of exercise therapy, including rock climbing, is that it comes with none of the potential adverse effects of pharmaceuticals – e.g. indigestion, nausea, diarrhoea, constipation, sweating, sudden heat stroke, swelling, and dry mouth – reported by up to 29% of patients treated [27]. It is also worth noting that 30–40% of patients treated with antidepressants appear to not respond to treatment [7, 28], further stressing the need for non-pharmaceutical treatment options and adjuvants.

The authors of this review have identified one previous systematic review [29], assessing the therapeutic effects of rock climbing on various conditions, of which depression was one subset. Comprising four randomized controlled trials (RCTs) on depression [30,31,32,33] it concluded that rock climbing positively affects depressiveness, but there was little evidence regarding the size of effect, and the strength of evidence was low. In addition, two previous narrative reviews were identified, one looking at general therapeutic effects of rock climbing, also suggesting that rock climbing reduces levels of depression [34], and one specifically looking at the effects on depression, also including four RCTs, also noting a positive effect of rock climbing on depression [35]. A few bachelor and master theses reviewing the subject were also identified, but none that included any additional RCTs. We therefore believe this new systematic review – including additional studies not included in previous reviews – and assessing the magnitude of effect – previously lacking – will further increase the knowledge on this topic. By following a rigorous scientific methodology, clearly describing the identified rock climbing interventions and their comparisons, synthesising the different findings, and assessing the overall certainty of evidence – also previously lacking – we believe this will result in higher levels of evidence and, consequently, meaningful clinical recommendations.

The aim of this review was to assess if rock climbing, in any form, can be used – and recommended – as an effective evidence-based treatment for depression. This general purpose was broken down into three specific research questions: Does rock climbing reduce symptoms of depression? Is the reduction, i.e. magnitude of effect, clinically relevant? Are the effects of rock climbing sustained long-term?

Methods

Protocol, registration and PRISMA statement

A protocol for this systematic review was registered in PROSPERO in January 2024 (PROSPERO ID: CRD42024468119). The review was conducted as per reported in the registered protocol without any deviations or violations. The review was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis statement (PRISMA) [36].

Eligibility criteria and PICO-framework

The PICO-framework was used to define eligibility criteria, with PICO standing for Population, Intervention, Comparison and Outcome [37, 38].

Population

Any individual diagnosed with, or showing symptoms of, at least moderate depression, measured by any validated depression rating scale, e.g. >20 points on BDI-II or > 20 points on MADRS. These specific cut-offs were used to standardize participant inclusion and minimize selection bias. The authors also contemplated excluding potential confounding factors, such as ongoing antidepressant treatment or ongoing CBT treatment, but decided against this after conducting preliminary exploratory searches and consulting a colleague in the field, since this would severely limit study population and generalisability.

Intervention

All forms of rock climbing, i.e. bouldering, sport climbing, traditional climbing, or any of its indoor equivalents. All intervention durations were eligible for inclusion.

Comparison

Any form of control was eligible for inclusion, e.g. other forms of exercise, talk/conversational therapy/counselling (including CBT and psychotherapy), antidepressants, treatment as usual, or waitlist. Presence of control group was a prerequisite for inclusion, in order to differentiate intervention results from regression toward the mean.

Outcome

Depression symptoms reduction or resolution measured with rating scales, e.g. Montgomery-Asberg Depression Rating Scale (MADRS) [39], Beck Depression Inventory-II (BDI-II) [40], Hospital Anxiety and Depression Scale (HADS) [41], or Symptom Checklist 90-R (SCL-90-R) [42].

Additional criteria

Only controlled trials were eligible for inclusion, but randomization was not a prerequisite (although lack of randomization could introduce bias, excluding non-randomized trials could reduce overall certainty of evidence by allowing for too few trials to be included). For the same reason, no limitation was set to sample size. To assert methodological rigor, only published peer-reviewed studies were considered for inclusion.

Literature search and identification of additional records

We conducted literature searches in PubMed (MEDLINE), Scopus, and Web of Science, in January 2024. A search strategy was developed for PubMed, then subsequently adapted to the other databases. Since no medical subject heading (MeSH) exist for rock climbing, we included the most proximate, i.e. “mountaineering”, reasoning it would cover rock climbing as well. The search strategy was designed with the help of a librarian at the Biomedical Library, University of Gothenburg, Sweden. Before finalizing the search strategy, we explored precision and recall using different search queries. The final search query consisted of combinations and variants of “climbing”, “bouldering”, “mountaineering”, “depression”, “Montgomery Asberg Depression Rating Scale”, “Beck Depression Inventory”, “Hospital Anxiety and Depression Scale”, and/or “Symptom Checklist 90-R”. We also explored searching the databases CINAHL, Cochrane Library, PEDro, and SPORTDiscus, but this did not produce any additional unique records. We also searched Google Scholar for grey literature, also in January 2024, using the same search terms, limiting results to the first 100 records, with relevance validated through peer-reviewed citations or author credentials, ensuring only credible sources were included. Finally, ClinicalTrials.gov was searched for ongoing trials. All databases were searched from inception up until January 2024. We did not apply any restrictions to language of publication date. Backward and forward citation searches of included studies, and identified previous reviews, were also performed, as well as consultation of content experts in the field. The search process was then repeated in January 2025, to make sure no new relevant studies had been published during time of drafting the manuscript. The complete search strategy, including all search terms, Boolean operators, field labels, filters, and adaptations across databases, is provided in Additional file 1, to ensure full transparency and replicability.

Screening process and study selection

Identified records were imported into Rayyan [43] for screening, after removing duplicates using EndNote [44]. All three authors (RL, LN and AL) screened titles and abstracts independently, according to eligible criteria. Records fulfilling PICO-criteria were marked as “Include” (i.e. going on to next stage of the screening process), and records not fulfilling PICO-criteria were marked as “Exclude”. If any uncertainty at this stage in the screening process, records were marked as “Include”, to be assessed during full-text screening. The authors were not blinded to trial identifiers such as authors’ and journals’ names. Before resolving any disagreements, interrater reliability between reviewers was calculated using Cohen’s kappa (κ) [45]. Disagreements were resolved through consensus discussions among all three authors. Full texts were then subsequently read independently, by all three authors. If the articles fulfilled PICO-criteria they were marked as “Include” (i.e. included in the review), otherwise as “Exclude” (with reasons for exclusion reported in a FLOW-diagram). Any remaining disagreements were again resolved using consensus discussions among all three authors. The screening and selection process was illustrated with a FLOW-diagram [36].

Data extraction

Data extraction included study design, participant demographics (including age and sex), intervention components for experiment and control group, outcome measures, and results. It was performed by two authors independently (RL and LN or AL). If necessary, corresponding authors were contacted for clarification of any uncertainties related to a study.

Study quality assessment

Two authors assessed methodological quality of included studies independently (RL and LN or AL), using the PEDro scale for quality. The PEDro scale is a 0–10-point scale assessing methodological quality of clinical trials [46], were < 4 points are considered poor quality, 4–5 points fair, 6–8 good, and 9–10 points excellent [47]. It has been shown to have both high reliability and high validity [48, 49]. After individual assessment, all articles were discussed among all three authors until consensus was reached. The results of the quality assessment were then used, together with other criteria, in the overall assessment of the certainty of evidence. A PEDro score of 6–10 points was considered low risk of bias, 4–5 points borderline risk, and < 4 points serious risk of bias. When multiple studies investigated the same outcome, the median PEDro score was used.

Data synthesis and analysis

Descriptive data were described with central tendencies (mean and median) and dispersion measures (standard deviation, SD, and interquartile range, IQR). Mean and SD were used for larger datasets, and median and IQR, as a more accurate measure, for smaller datasets.

When probability values (p-values) for between-group differences were missing, these were calculated using MedCalc [50] based on mean, SD, and sample sizes. When standard deviations (SDs) were missing (necessary for the funnel plot), these were calculated using the RevMan Calculator [51].

When assessing intervention effectiveness (i.e. magnitude of effect), a within-group change (Δ) above the minimal clinically important difference (MCID) was considered clinically relevant. When assessing intervention superiority, a between-group difference above MCID was considered clinically relevant. For MADRS, a difference of five points (8%) has been considered as MCID [52] and was used for this review (NB: this was two points higher than the cut-off used in some of the included studies, e.g. Luttenberger et al. [53]). For BDI-II a MCID of between five and 11 points has been suggested [54, 55]. For this review a six-point difference was considered clinically relevant, since this both corresponded with the cut-off used in the included study [32] as well as the general 10% suggested by Cohen [56]. MCID for Patient Health Questionnaire-9 (PHQ-9) has been suggested to be between three (11%) and five (19%) points [57, 58]. For this review, a MCID of three points was considered clinically relevant, since this best corresponded with the other MCIDs. For Symptom Checklist-90-Revised (SCL-90-R), and its subscale depression, a MCID could not be identified, and a mean difference of > 10%, corresponding to 4 points, was considered clinically relevant [56]. For 5-point Likert scales like the Positive and Negative Affect Schedule (PANAS), a mean difference of > 10%, corresponding to 0.5 points, was considered clinically relevant [56]. These thresholds reflect changes likely to be perceived as meaningful by patients, beyond statistical significance.

Due to the large clinical heterogeneity of included studies, regarding their interventions (i.e. type and duration), and especially their comparisons (which were as versatile as to include waitlist, relaxation and physical exercises, and CBT-treatment), a meta-analysis could not be merited. Instead, the results were presented using a narrative synthesis, organised according to intervention duration and their respective comparisons. We used a textual approach, grouping trials based on their intervention length (duration), type of intervention, and comparisons. When this resulted in more than one trial exploring the same outcome, mean change (Δ) and mean between-group difference were calculated, as opposed to simple vote counting. These outcomes were then assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.

All analyses were performed using IBM SPSS (v. 28.0) [59], RevMan Calculator [51], or MedCalc [50].

Certainty of evidence (GRADE) assessment

We assessed the certainty of evidence using GRADE, with the following criteria: risk of bias, inconsistency, indirectness, imprecision, and reporting bias [60, 61]. Results from randomised trials were initially assigned a certainty level of four (high), then rated down half a point (borderline risk), one point (serious risk) or two points (very serious risk) if we detected issues with risk of bias, inconsistency, indirectness, or imprecision (mainly due to low number of studies) respectively. Results based on non-randomised trials were initially assigned a certainty level of two (low) but were not rated down further for risk of bias. Issues with inconsistency, indirectness, or imprecision were rated down by half a point, one point, or two points if detected. Publication bias was assessed for all outcomes simultaneously, using a funnel-plot [62] (see Additional file 2 for details), although due to the low number of included studies (n = 7) there is a high risk of this test being underpowered [63]. We also searched ClinicalTrials.gov for study protocols of unpublished studies (only identifying study protocols for published or ongoing studies), assessed the funding sources (see Additional file 3 for details) and considered the extensive search strategy (including grey literature) and the intervention of interest (non-proprietary) when assessing publication bias.

We saw no reason to rate up the level of evidence neither due to large effect sizes, dose-response effects nor effects of residual confounding factors.

The overall certainty of evidence for each outcome was then assessed as either high (4/4: a very good indication of the likely effect), moderate (3/4: a good indication of the likely effect), low (2/4: some indication of the likely effect), or very low certainty (1/4: not a reliable indication of the likely effect).

Results

Search results

The search process generated 3,364 records, of which 1,933 remained after removing duplicates. A total of 1,902 records were excluded during screening of titles and abstracts. Reviewer agreement during title and abstract screening ranged from almost perfect (RL, AL; κ = 0.81, p < 0.01) to substantial (LN, AL; κ = 0.80, p < 0.01) and moderate (RL, LN; κ = 0.69, p < 0.01). After screening the remaining full texts, an additional 21 records were excluded. Reviewer agreement during full text screening ranged from substantial (RL, AL; κ = 0.75, p < 0.01) to moderate (LN, AL; κ = 0.53, p = 0.01) and fair (RL, LN; κ = 0.31, p = 0.10). Backward and forward citation searches of included studies did not yield any additional relevant records. In total, seven studies, reported in ten articles [30,31,32,33, 53, 64,65,66,67,68] were included. Three articles [65, 66, 68] reported on the same sample/intervention/comparison as an already included study [53], but reported on different outcomes not relevant for this review. The screening and selection process is illustrated in Fig. 1.

Fig. 1
Fig. 1
Full size image

PRISMA 2020 flow diagram of the search and inclusion process. Source: Page et al. [36]. This work is licensed under CC BY 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/

Study quality assessment

Included studies had a PEDro methodological quality score of between four (fair) and eight (good) out of ten. Median quality score of included studies was six (IQR 5–7), i.e. good quality. See Table 3 for individual study PEDro scores, and Additional file 4 for a detailed breakdown of the individual PEDro scores. None of the included studies blinded neither patients nor therapists, but two studies [30, 53] blinded the assessors. Three studies had a dropout rate just above the 15% cut-off (range 16–18%) [33, 53, 64], two studies [32, 33] did not use intention-to-treat analysis, one study [31] did not use concealed or random allocation, and in one study [31] groups differed at baseline for one main outcome.

Table 3 Summary of included studies (n = 7, reported in 10 articles)
Full size table

Characteristics of included studies

Included studies were conducted between 2015 and 2023, nine in Germany and one in Austria [64]. Of the studies conducted in Germany, all but one [31] were conducted by the same team of researchers. Five of the studies [30, 53, 65, 66, 68] also reported on the same sample and intervention, but focused on different comparisons and outcomes. Studies were funded either by a charitable foundation [30, 53] or a university hospital [32, 33, 67], and for the climbing entrance fees directly by the climbing gym [30, 53]. Two studies [31, 64] did not receive any funding. See Additional file 3 for details of the funding sources.

A total of 471 participants were included in the studies (participants reported in multiple articles only counted once), all with moderate depression according to BDI-II/PHQ-9/WHO-5 Well-Being Index (WHO-5) or diagnosed with depression/major depressive disorder/bipolar disorder by psychiatrist, according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-4). Mean age of participants across the studies were 43 years, with 270 being female and 197 male (due to dropout, sex of four participants were not reported). Interventions consisted of a combination of indoor bouldering and mindfulness exercise, named “Bouldering Psychotherapy” (“BPT”) in the articles [30, 32, 33, 53, 67] or top-rope climbing on an artificial climbing wall [31, 64]. Intervention durations ranged from a single 2.5-hour session [31] to 1.5–3-hour sessions, 1–2 times a week, for 4–10 weeks. Interventions were conducted by either psychologists, psychotherapists, psychotherapists-in-training, professionals in psychology/sport science/training therapy/medicine, or registered nurses (RNs), with several years of experience in rock climbing and bouldering. Comparisons consisted of Nordic walking [64], supervised home-exercise training and psychoeducational material [30], group CBT led by licensed CBT psychotherapists/psychotherapists-in-training [53], movies on anxiety/post-traumatic stress disorder (PTSD) and guided group dialogue [64], group relaxation exercise and “feelings discussion” led by psychologist [31], or waitlist [32, 33, 67]. Outcomes on depression symptoms reduction were measured with BDI-II [32, 33, 64, 67], MADRS [30, 53], SCL-90-R subscale depression [32, 33], PHQ-9 [30, 53], PANAS [31], or authors own 5-point Likert scale for depressiveness [31]. Detailed characteristics of included articles are presented in Table 3.

Acute (hours) effects of rock climbing on depression

One controlled clinical trial of fair quality [31] examined the acute effects of rock climbing.

A single session of indoor top-rope climbing showed a statistically significant reduction of depressive symptoms exceeding MCID (> 10% on PANAS), i.e. clinically relevant. PANAS 1.0 (95% CI 0.6, 1.3); −0.9 (95% CI −1.2, −0.5); “depressiveness” −0.9 (95% CI −1.3, −0.5). The effects were greater than those of a group relaxation exercise and “feelings discussion”, between-group difference PANAS 1.4 (95% CI 1.0, 1.7); 0.5 (95% CI 0.1, 0.9); “depressiveness” 1.1 (95% CI 0.6, 1.5). Very low certainty evidence (downgraded for very serious risk of bias due to non-randomisation, and very serious imprecision due to only one study).

Short-term (4–10 weeks) effects of rock climbing on depression

Five studies of good quality [25, 32, 53, 64, 67], and one of fair [33], examined the short-term effects of rock climbing.

Eight to ten weeks of indoor bouldering (in combination with mindfulness exercises) led to a statistically significant reduction of depressive symptoms from moderate (e.g. MADRS: 20–34) to mild (e.g. MADRS: 7–19), exceeding MCID (e.g. 5 points on MADRS). MADRS mean − 8.3 (SD 10.1–10.4), p < 0.05; PHQ-9 mean − 4.7 (SD 6.1–6.3), p < 0.05; BDI-II mean − 7.3 (SD 5.6–9.4), p = 0.01–0.09; SCL-90-R mean − 5.7 (SD 4.4–9.1), p = 0.04. It was also significantly better than waitlist/no intervention, between-group difference BDI-II mean 5.1, p = 0.01–0.09; SCL-90-R mean 3.8, p < 0.05, and in two studies showed a between-group differences exceeding MCID. High certainty evidence.

Indoor bouldering (in combination with mindfulness exercises) was also significantly better, with a difference exceeding MCID (5 points on MADRS), than supervised home-exercise training, between-group difference MADRS 5.4, p < 0.01; PHQ-9 2.1, p = 0.04, but not better than group CBT-treatment, between-group difference MADRS 2.1 (95% CI − 1.0, 5.1); PHQ-9 1.2 (95% CI − 0.7, 3.0). Low certainty evidence (downgraded for very serious imprecision due to only one study).

Four weeks of top-rope climbing on artificial climbing wall led to a statistically significant reduction of depressive symptoms from moderate (BDI-II: 20–28) to mild (BDI-II: 14–19), exceeding MCID (6 points on BDI-II). BDI-II −7.1p (95% CI −13.8, −0.4), but was not better than neither Nordic walking, BDI-II 2.1, p = 0.72, nor anxiety/PTSD-movies and guided group dialogue, BDI-II 5.0, p = 0.72. Low certainty evidence (downgraded for very serious imprecision due to only one study).

Long-term (6–12 months) effects of rock climbing on depression

Four studies of good quality [32, 53, 64, 67], and one of fair [33], examined the long-term effects of rock climbing.

Eight to ten weeks of indoor bouldering (in combination with mindfulness exercises) showed a statistically significant sustained reduction of depressive symptoms from moderate (e.g. MADRS: 20–34) to mild (e.g. MADRS: 7–19) after six and twelve months. MADRS − 10.0p (SD 10.5), p < 0.01; PHQ-9 −5.6 (SD 6.0), p < 0.01; BDI-II −4.4p, p = 0.01. High certainty evidence.

There was no long-term difference in comparison to group CBT-treatment. MADRS 0.7 (95% CI − 2.6, 3.9); PHQ-9 0.7 (95% CI − 1.1, 2.6). Low certainty evidence (downgraded for very serious imprecision due to only one study).

Four weeks of top-rope climbing on artificial climbing wall showed a small, sustained effect after six months. BDI-II −4.8p, p < 0.1, but was not better than neither Nordic walking, BDI-II 2.6p, p = 0.72, nor anxiety/PTSD-movies and guided group dialogue, BDI-II 2.0p, p = 0.72. Low certainty evidence (downgraded for very serious imprecision due to only one study).

Negative outcomes of rock climbing

No adverse effects or injuries were reported in any of the studies, neither for the climbing interventions nor for the comparisons.

Summary of findings (GRADE) for effects of rock climbing on depression

A summary of the findings and certainty of evidence according to GRADE are presented in Table 4.

Table 4 Summary of findings (GRADE) for effects of rock climbing on depression symptoms
Full size table

Discussion

Based on six RCTs and one controlled clinical trial, all with a generally good methodological quality, and including a total of 471 participants, high certainty evidence suggests that indoor bouldering (combined with mindfulness exercises) is effective in reducing symptoms of depression from moderate to mild, compared to no intervention. High certainty evidence also suggests that these effects are sustained long-term. Low certainty evidence suggests that four weeks of top-rope climbing reduces symptoms of depression from moderate to mild, are sustained long-term, but may not be clinically relevant. Low certainty evidence suggests that indoor bouldering (combined with mindfulness exercises) is more effective than a home exercise program, but not better than group CBT, and that top-rope climbing is not better than Nordic walking or anxiety/PTSD-movies together with guided group dialogue. Very low certainty evidence suggests that a single top-rope climbing session may be better than a group relaxation exercise and “feelings discussion”. The lower certainty level for certain outcomes are primarily due to insufficient trials investigating these comparisons, and for the acute effects also due to lack of randomisation.

Interpretation and comparison with previous studies and reviews

The above results are all in accordance with those reported elsewhere, i.e. a German master thesis (Huffer, 2023, personal communication), a German bachelor thesis (Schiller, 2019, personal communication), a Spanish master thesis [69] and on a Swiss conference [70], identified in our search process or through correspondence with other authors, but not included in our review due to methodological reasons. Our results are also in accordance with the previous review by Zieliński et al. [35] and the reviews by Gassner et al. [29] and Liu et al. [34], the latter two looking not only on effects on depression, but on other therapeutic effects of rock climbing as well. The previous reviews based their findings on some of the same trials here included, but our review is the first to have included a total of seven trials, investigating the effects of rock climbing on depression, and the first to use a GRADE-approach, thus enhancing confidence in the findings. By clearly defining the population (i.e. individuals with at least moderate depression), focusing solely on depression as main outcome, and using MCID thresholds, we believe our review provide strong, depression-specific, evidence. Also, by using a time-based outcome structure, previously lacking, we believe onset and long-term sustainability of effects are more easily appreciated. The use and structured reporting of MCID also allows for a more direct clinical applicability of results. We therefore believe this review provides the most clinically relevant synthesis of the current evidence on rock climbing for depression to date.

Although no injuries were reported in any of the included studies, prior epidemiological research suggests that the risk of injury in climbing is present, with up to 44% of participants sustaining an overuse injury throughout their indoor climbing career, with bouldering leading to more injuries than top-rope climbing [71]. The risk of injury has been assessed to be 0.2 injuries per 1,000 h of rock climbing [72]. In comparison, this is considerably lower than reported in other sports, were number of injuries have been reported to be between 0.4 (swimming) and 7.2 (soccer) per 1,000 h of playing and training, and with 44% of participants sustaining an injury in the last year [73]. A possible explanation for the absence of injuries in the included rock climbing studies is that the accumulated intervention durations never exceeded 20–24 h (two hours once a week for 10 weeks or three hours once a week for eight weeks), which is simply not enough time for an overuse injury to manifest.

It should also be stressed that – while the climbing interventions were not always superior to their comparisons (anxiety/PTSD-movies and guided group dialogue, or CBT) – rock climbing, unlike those interventions, improves physical fitness [74]. Although physical fitness was not an outcome of this review, it is known through a plethora of other research to be associated with an overall reduction in all-cause mortality [75], and clearly has an intrinsic value of its own. Rock climbing also comes with none of the side effects of antidepressants [27]. So, while superiority evidence is lacking, and the strongest evidence comes from adjunctive therapies, all evidence points in the same direction, i.e. lowered symptoms of depression from indoor rock climbing. Also considering that no adverse effects were reported, and that rock climbing improves physical fitness, we believe that indoor rock climbing (especially bouldering, with added mindfulness exercises) can be recommended as an adjunctive therapy for mild to moderate depression, due to its unique combination of physical and psychological elements.

Generalisability

Of the studies identified and included in our review, only two [31, 64] investigated rock climbing/top-rope climbing in isolation, while the other five [30, 32, 33, 53, 67] used an adjunctive intervention protocol, combining indoor bouldering with mindfulness exercises. Unfortunately, this makes it difficult to assess whether the intervention effects are due to indoor bouldering, mindfulness exercise, or the combination of the two. The conclusions of this review are based on the assumption that the therapeutic effects result from a combination of the two interventions, but it is impossible to determine the extent to which each contributed to the total. As proposed in the introduction, by Wheatley [26], it is also possible that rock climbing alone may induce a mindfulness-like state, potentially reducing the added value of a separate mindfulness exercise. Therefore, dedicating more time to rock climbing itself, rather than dividing time between rock climbing and specific mindfulness exercises, may yield a greater synergistic effect. Future research should therefore separate rock climbing interventions from any adjunctive therapies, when assessing its effects on depressive symptoms, to help identify which components (physical, cognitive, or mindfulness-related) contribute the most to the observed benefits.

For within-group comparisons over time, and comparisons to waitlist, several studies investigated these effects. However, for the active comparisons – such as Nordic walking, anxiety/PTSD-movies and guided group dialogue, supervised home-exercise training, group relaxation exercises and “feelings discussion”, and group CBT – only one study investigated each comparison, respectively. This limits the certainty of evidence for these outcomes. Also, no study has to date compared rock climbing/bouldering to pharmaceuticals (e.g. SSRIs/SNRIs/TCAs) or aerobic group exercise, which are closer to a gold standard therapy for depression [11], further limiting the confidence in potential superiority effects.

In all included studies, trained medical personnel delivered the rock climbing interventions. This was also seen in e.g. Engbert and Weber [19], Schinhan et al. [21], and Wheatley [26], although in Wheatley, the bouldering therapy was ultimately performed unsupervised, after an introduction and signing of a waiver. If this is a necessary prerequisite, it makes the intervention of rock climbing rather inaccessible in a real-world context, as it would require instructors with both a medical and a rock climbing background. If, by comparison, the same or somewhat similar results could be obtained under guidance of a climbing instructor without medical training, such an intervention would be much more accessible. We believe that, as long as the rock climbing intervention is kept separate from e.g. mindfulness exercise, this is possible. It would therefore be of interest for future research in the field to explore the delivery of climbing interventions by non-medical personnel.

Mean age of participant included in our analysis were 43 years, which is higher than that of median age for onset of depression, which has been reported to be 30 years of age [76]. It is also considerably higher than the age at which people generally participate in sports, where one study [77] found 64% of sports-participants being below 20 years of age. It is therefore possible that the age of the participants may limit generalizability to a younger population, and caution should be applied when extrapolating the results to younger populations. The influence of age should be further explored in future studies by including a younger population sample. All studies, apart from Kleinstäuber et al., [31] had also explicitly excluded participants with physical/somatic disorders/inabilities. The absence of participants with somatic comorbidities may therefore limit generalizability to patients with co-occurring physical conditions. Future studies should make an effort to include such participants to enhance applicability.

The rock climbing interventions in all included studies were also administered as group therapy, consisting of groups of between eight and thirteen participants, whereas the control intervention was administered either in similar groups (Nordic walking; anxiety/PTSD-movies and guided group dialogue; group relaxation exercise and “feelings discussion”; group CBT) or individual (home exercise) or no treatment. When climbing was compared to no intervention, or individual home-exercise, it proved superior. When compared to other group interventions, it was generally no difference, i.e. compared to Nordic walking, anxiety/PTSD-movies and guided group dialogue, or group CBT. Only compared to group relaxation exercise and “feelings discussion” did rock climbing prove superior, but certainty of this evidence was very low. There is, therefore, a chance that the reported effects are not due to the specific interventions, but rather to the fact that they were performed in groups. This confounder is also discussed by Karg et al. [30] and Luttenberger et al. [32], and it further emphasizes the NICE guidelines [11], highlighting that group exercise has the strongest evidence-based foundation when it comes to reducing symptoms of depression. This is likely due to the established fact that feelings of belonging, i.e. group cohesiveness, has therapeutic effects of its own [78]. It is therefore recommended that future studies on the effect of rock climbing/bouldering on depression should evaluate both individual as well as group settings, comparing individual versus group rock climbing interventions, or adjusting for group cohesiveness using validated scales (e.g. Group Environment Questionnaire), to isolate the specific effects of rock climbing. Because maybe it is not what we do, but how we do it, that matters most.

Limitations and methodological considerations

Main methodological caveats of included studies were the inability to blind patients and therapists, which is assumed to be associated with a higher risk of bias. This is, for obvious methodological reasons, the standard within physical therapy and exercise intervention research [79]. However, the risk of bias associated with lack of blinding, when comparing two active interventions, should not be exaggerated. A meta-epidemiological study [80], on the effects of blinding in physical therapy interventions, found no significant inflation of effect sizes when comparing trials that blinded participants with those that did not.

Of note is also that five of the seven included studies [30, 32, 33, 53, 67] were conducted by the same team of researchers, and two of these [30, 53] reported on the same intervention group, which raises the possibility of data redundancy. While publication bias was judged unlikely, it would be advantageous to see the results replicated by other researchers, using different population-samples.

Agreement between reviewers in our review during full text screening ranged from fair to moderate, which were likely due to two factors. Firstly, an a priori cut-off for actual depression on the various rating scales, e.g. MADRS, BDI-II, HADS, SCL-90-R, had not been determined, resulting in this having to be decided during the screening process. A decision was then made to only include trials where participants had at least symptoms of moderate depression, or an actual diagnosis of depression. This was done to avoid a ceiling effect, which would arise if participants with mild or no depression were included. Secondly, several articles did not clearly describe their interventions, necessitating discussions on their potential inclusion. In the end, only articles which clearly described the interventions were included, in order not to confound the results.

The decision not to control for concomitant treatments, in order to increase sample size and generalizability, might have negatively affected internal validity. Although, a post hoc assessment of pharmaceutical treatments or additional psychotherapy between study groups showed no significant differences, except for one study (25), where the climbing group received more additional psychotherapy than the control exercise group, thereby possibly affecting the results. The large heterogeneity in study settings of included studies also prevented pooling of results in a meta analysis and precluded any reliable evidence regarding superiority effects. The use of PEDro to assess risk of bias is also somewhat indirect, as methodological scores are translated to risk of bias scores, but direct comparisons to e.g. the Cochrane risk of bias 2-tool show similar items/domains [81], and we have no reason to believe this have affected the outcome.

Strengths of the review

A strength of this review was the extensive and comprehensive literature search, with over 1,800 records screened for inclusion, including grey literature, with no exclusion being made for neither date nor language of publication. This makes it unlikely that any available records were missed. The scrutiny of reference lists also produced no new relevant records, which can be regarded as an indicator of an exhaustive search strategy [82].

Only controlled clinical trials, of generally good methodological quality, were included, with no inconsistencies or indirectness detected. This increases the robustness of the results. In addition, the use of the GRADE approach allowed us to place great confidence in the certainty of our findings. The consistent use of MCID thresholds also increases clinical usability. To our knowledge, this is the first systematic review to combine these methodological frameworks in evaluating climbing-based interventions for depression.

Conclusions

This systematic review found that indoor rock climbing – particularly bouldering combined with mindfulness exercises – significantly reduces depressive symptoms in adults with moderate depression, with effects sustained for up to 12 months. Compared to waitlist and individually performed home-based exercises, between-group differences also exceeded MCID thresholds for clinical relevance. No adverse events were reported across studies. The certainty of evidence ranged from high (for short- and long-term effects of bouldering versus waitlist) to very low (for acute effects), suggesting that indoor rock climbing may be a promising and safe adjuvant intervention for depression. Further high-quality RCTs are needed to isolate intervention effects from potential confounding by group dynamics and age, and to compare climbing directly with established treatments such as antidepressants and aerobic exercise.